-
Notifications
You must be signed in to change notification settings - Fork 4
/
Example_script.py
332 lines (237 loc) · 9.76 KB
/
Example_script.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
import FragMent
import numpy
import matplotlib.pyplot as plt
from scipy.stats import iqr
from astropy.io import fits
from matplotlib import gridspec
### Load the positions of the cores
pos = numpy.loadtxt("core_positions.dat")
### Load the spine of the filaments
spine = numpy.loadtxt("spine.dat")
### Define a boundary for placing random cores
xmin = 1.994
xmax = 1.996
ymin = 0.0
ymax = 4.0
boundary = numpy.array([xmin,xmax,ymin,ymax])
### Perform a nearest neighbour test on the data
nn_seps = FragMent.NNS(pos)
### Statistical significance of the results
nruns = 1000
sep_limit = 0.06
p_median, p_mean = FragMent.Stat_Sig(nn_seps, "NNS", boundary, nruns, sep_limit)
ks_stat, p_ks = FragMent.KS_test(nn_seps, "NNS", boundary, nruns, sep_limit)
ad_stat, crit_vals, p_ad = FragMent.AD_test(nn_seps, "NNS", boundary, nruns, sep_limit)
### Report results
print(" ")
print(" ")
print("######## Nearest neighbour results ########")
print("The median and interquartile range of the distribution are: ", numpy.median(nn_seps), iqr(nn_seps))
print("The p-value using the median-interquartile range NHT is : ", p_median)
print("The mean and standard deviation of the distribution are : ", numpy.mean(nn_seps), numpy.std(nn_seps))
print("The p-value using the mean-standard deviation NHT is : ", p_mean)
print("The p-values from the K-S and A-D test are : ", p_ks, p_ad)
### Perform a minimum spanning test on the data
mst_seps, mst = FragMent.MST(pos)
### Statistical significance of the results
p_median, p_mean = FragMent.Stat_Sig(mst_seps, "MST", boundary, nruns, sep_limit)
ks_stat, p_ks = FragMent.KS_test(mst_seps, "MST", boundary, nruns, sep_limit)
ad_stat, crit_vals, p_ad = FragMent.AD_test(mst_seps, "MST", boundary, nruns, sep_limit)
### Report results
print(" ")
print(" ")
print("######## Minimum spanning tree results ########")
print("The median and interquartile range of the distribution are: ", numpy.median(mst_seps), iqr(mst_seps))
print("The p-value using the median-interquartile range NHT is : ", p_median)
print("The mean and standard deviation of the distribution are : ", numpy.mean(mst_seps), numpy.std(mst_seps))
print("The p-value using the mean-standard deviation NHT are : ", p_mean)
print("The p-values from the K-S and A-D test are : ", p_ks, p_ad)
### Perform a Nth nearest neighbour test on the data
NNN_sep = FragMent.NNNS(pos)
### Reduce the separation 2d-array into an average and a width
mean_seps = numpy.mean(NNN_sep, axis=0)
std_seps = numpy.std(NNN_sep, axis=0)
median_seps = numpy.median(NNN_sep, axis=0)
iqr_seps = iqr(NNN_sep, axis=0)
mean_seps = mean_seps[1:]
std_seps = std_seps[1:]
median_seps = median_seps[1:]
iqr_seps = iqr_seps[1:]
### Plot results
plt.figure(1)
plt.plot(mean_seps,std_seps,"b",label="Mean")
plt.plot(median_seps,iqr_seps,"r",label="Median")
plt.legend(loc="best")
plt.xlabel("Average separation")
plt.ylabel("Average width")
plt.savefig("NNN_results.png")
print(" ")
print(" ")
print("######## Nth nearest neighbour test ########")
print("The results are plotted in the file NNN_results.png")
### Determine the two-point correlation function
### Calculate the exact two point correlation function
nruns = 1000
lower_lim = 0.06
sep, w = FragMent.TwoPoint(pos,nruns,boundary,lower_lim)
### Calculate the approximate two point correlation function
op_crit = 1e-2
sep, wa, sig_w = FragMent.ApproxTwoPoint(pos,nruns,boundary,lower_lim,op_crit)
### Plot results
plt.figure(2)
plt.plot(sep,w,"b",label="Exact")
plt.plot(sep,wa,"k--",label="Approx")
plt.fill_between(sep,wa-sig_w,wa+sig_w,alpha=0.5)
plt.xlabel("Separation")
plt.ylabel("Two point correlation function")
plt.legend(loc="best")
plt.savefig("Two_point_results.png")
print(" ")
print(" ")
print("######## Two point correlation function test ########")
print("The results are plotted in the file Two_point_results.png")
### Determine the power spectrum of the spine data
k,power = FragMent.FT_Spine(spine)
### Remove the k=0 mode
k = k[1:]
power = power[1:]
### Plot results
plt.figure(3)
plt.plot(k,power/power[0],"b")
plt.xlabel("Wavenumber")
plt.ylabel("Normalised power spectrum")
plt.savefig("FT_results.png")
print(" ")
print(" ")
print("######## Power spectrum test ########")
print("The results are plotted in the file FT_results.png")
### Frequentist model selection
### Place boundaries on the five parameters of the two-tier fragmentation model (the first 4 entries are the boundaries for the single-tier model)
bound = numpy.zeros(10)
### Boundaries for the mean of the first Gaussian
bound[0] = 0.0
bound[1] = 1.0
### Boundaries for the standard deviation of the first Gaussian
bound[2] = 1e-99
bound[3] = 0.5
### Boundaries for the mean of the second Gaussian
bound[4] = 0.0
bound[5] = 1.0
### Boundaries for the standard deviation of the second Gaussian
bound[6] = 1e-99
bound[7] = 0.5
### Boundaries for the ratio of the amplitude of the two Gaussians
bound[8] = 0.0
bound[9] = 10.0
### Setup the MCMC parameter array
mcmc_param = numpy.zeros(3,dtype=int)
mcmc_param[0] = 100 ### Number of walkers
mcmc_param[1] = 200 ### Number of burn steps for each walker
mcmc_param[2] = 1000 ### Number of actual steps for each walker
### Run the model selection
model_one, theta_one, model_two, theta_two = FragMent.Freq_model_selection(mst_seps,bound,mcmc_param)
### Unpack outputs
aic_one = model_one[0]
weight_one = model_one[1]
aic_two = model_two[0]
weight_two = model_two[1]
### Report results
print(" ")
print(" ")
print("######## Frequentist model selection results ########")
if(aic_one < aic_two):
print("Single-tier fragmentation is preferred")
print("The Akaike weight for the model is: ", weight_one)
print("This gives an evidence ratio of : ", weight_one/weight_two)
print("The fit parameters are : ", theta_one)
if(aic_two< aic_one):
print("Two-tier fragmentation is preferred")
print("The Akaike weight for the model is: ", weight_two)
print("This gives an evidence ratio of : ", weight_two/weight_one)
print("The fit parameters are : ", theta_two)
### Bayesian model selection
### Declare the priors for model one and model two
### Both priors are flat, i.e. all model parameters have the same probability
def prior_one(params, bound):
mu1, sig1 = params
term1 = 1.0/(bound[1]-bound[0])
term2 = 1.0/(bound[3]-bound[2])
return term1 * term2
def prior_two(params, bound):
mu1, sig1, mu2, sig2, ratio = params
term1 = 1.0/(bound[1]-bound[0])
term2 = 1.0/(bound[3]-bound[2])
term3 = 1.0/(bound[5]-bound[4])
term4 = 1.0/(bound[7]-bound[6])
term5 = 1.0/(bound[9]-bound[8])
return term1*term2*term3*term4*term5
### Calculate the posterior for each model, these can be plotting using corner
walkers_one, post_one, walkers_two, post_two = FragMent.Posterior_of_models(mst_seps,bound,mcmc_param,prior_one, prior_two)
### Calculate the evidence of each model
n=15
evi_one, evi_two = FragMent.Evidence_explicit(mst_seps,bound,prior_one, prior_two,n)
### Report results
print(" ")
print(" ")
print("######## Bayesian model selection results ########")
if(evi_one > evi_two):
print("Single-tier fragmentation is preferred")
print("The Bayes factor for in favour of model one is : ", evi_one/evi_two)
print("The most likely parameters from the posterior are: ", walkers_one[numpy.argmax(post_one)])
if(evi_two > evi_one):
print("Two-tier fragmentation is preferred")
print("The Bayes factor for in favour of model two is : ", evi_two/evi_one)
print("The most likely parameters from the posterior are: ", walkers_two[numpy.argmax(post_two)])
### An example of straightening a filament and the accompanying cores
### Load the example filament map and spine
Map = fits.getdata("./Curved_filament.fits")
spine = numpy.loadtxt("./Curved_filament_spine.dat")
### Place some `cores' close to the spine
pos = numpy.zeros((4,2))
pos[0,0] = spine[5,0]
pos[0,1] = spine[5,1] + 1
pos[1,0] = spine[17,0] + 3
pos[1,1] = spine[17,1] + 1
pos[2,0] = spine[60,0]
pos[2,1] = spine[60,1]
pos[3,0] = spine[100,0] + 2
pos[3,1] = spine[100,1] - 3
### Map the cores to the spine
cpos = FragMent.Map_cores(spine,pos,10)
### Straighten the filament using the interpolation and weighted average methods
s,l,r = FragMent.Straighten_filament_interp(spine,Map.T,90,30,10)
s2,l,r = FragMent.Straighten_filament_weight(spine,Map.T,90,30,10,0.5)
### Get limits for the plots
miny = numpy.amin(l)
maxy = numpy.amax(l)
minx = numpy.amin(r)
maxx = numpy.amax(r)
maxc = numpy.amax(Map)
minc = numpy.amin(Map)
### Plot the results
fig = plt.figure(4,figsize=(10,8))
gs = gridspec.GridSpec(2,4)
ax1 = fig.add_subplot(gs[:,:2])
i = ax1.imshow(Map,origin=0,interpolation="gaussian", vmax = maxc, vmin = minc , extent=[0,200,0,200])
ax1.plot(spine[:,0],spine[:,1],"k",linewidth="2")
ax1.plot(pos[:,0],pos[:,1],"ro")
plt.setp(ax1,yticks=[40,80,120,160],yticklabels=[40,80,120,160],xticks=[40,80,120,160],xticklabels=[40,80,120,160])
plt.ylabel("y")
plt.xlabel("x")
ax2 = fig.add_subplot(gs[:,2])
ax2.imshow(s.T,origin="lower",interpolation="gaussian", vmax = maxc, vmin = minc , extent=[minx,maxx,miny,maxy])
ax2.plot(cpos[:,0],cpos[:,1],"ro")
plt.setp(ax2,yticks=[],yticklabels=[],xticks=[-15,0,15],xticklabels=[-15,0,15])
plt.xlabel("Radius")
ax3 = fig.add_subplot(gs[:,3])
ax3.imshow(s2.T,origin="lower",interpolation="gaussian", vmax = maxc, vmin = minc , extent=[minx,maxx,miny,maxy])
ax3.plot(cpos[:,0],cpos[:,1],"ro")
plt.setp(ax3,yticks=[20,40,60,80,100],yticklabels=[20,40,60,80,100],xticks=[-15,0,15],xticklabels=[-15,0,15])
plt.ylabel("Longitudinal axis")
plt.xlabel("Radius")
ax3.yaxis.tick_right()
ax3.yaxis.set_label_position("right")
cbaxes = fig.add_axes([0.1, 0.1, 0.8,0.05])
cb = plt.colorbar(i,label="Intensity", cax = cbaxes,orientation="horizontal")
fig.subplots_adjust(left=0.11, bottom=0.18, right=0.89, top=0.93, wspace=0.25, hspace=0.1)
plt.savefig("Straight_filament.png")